In the modern era where computers are taking over the world, the vehicles are focused to be fully computerized to give human driver a safe and relaxed driving. As humans naturally make mistakes, autonomous vehicles are mainly focused on errorless driving to make the journey safe. The goal is to use computer vision as well as machine learning algorithms to let the vehicles drive themselves. After capturing each frame from the simulation platform, computer vision recognizes the lanes and barriers. The driving vehicle will proceed based on decisions made after autonomously identifying barriers and traffic signals. Euro Truck Simulator is used to test the model against the simulation platform. Several optimizers, including Adam, Adagrad, Adamax, Adadelta, and SGD, as well as activation functions, including sigmoid, exponential linear unit (ELU), and rectified linear unit (RELU), are used to train the model. The observations show that the greater accuracy of 0.9889 is achieved with the ELU activation function and the Adam optimizer.

1.
S.
Behere
and
M.
Torngren
, “
A functional reference architecture for autonomous driving
,”
Information and Software Technology
,
0
,
1
15
(
2015
).
2.
R.
Girshick
, “
Fast r-cnn
,”
International Conference on Computer Vision (ICCV)
, (
2015
).
3.
Veres
,
Sandor
&
Molnar
,
Levente
&
Lincoln
,
Nick
&
Morice
,
Colin
, “
Autonomousvehicle control systems - A review of decision making
”,
IMechE Journal of Systems and Control
, (
2010
).
4.
E.
Yurtsever
,
J.
Lambert
,
A.
Carballo
,
K.
Takeda
, “
A Survey of Autonomous Driving: Common Practices and Emerging Technologies
”,
IEEE Access
,
58443
58469
(
2020
).
5.
Liang
,
Xiaodan
&
Wang
,
Tairui
&
Yang
,
Luona
&
Xing
,
Eric
, “
CIRL: Controllable Imitative Reinforcement Learning for Vision-Based Self-driving
”,
15th European Conference, Munich, Germany
, (
2018
).
6.
Q.
Zhang
, &
T.
Du
, “
Self-driving scale car trained by Deep Reinforcement Learning‖, ArXiv
, (
2019
).
This content is only available via PDF.
You do not currently have access to this content.